Chronic pain after bilateral thoracotomy in lung transplant patients

2.1 Pain assessment

For each side of the thorax, patients completed the Neuropathic Pain Symptom Inventory (NPSI), which contains 12 questions about neuropathic pain characteristics [17]. Patients also answered questions about the frequency of pain in and around the scar during the last week (present constantly, every day, with days interval, no pain) and intensity of pain during the last week and at present at rest, using a numeric rating scale (NRS) from 0 to 10. In addition, without distinguishing between the two sides of the thorax, patients answered questions about the impact of pain in and around the scar on daily activities (no impact, light, moderate, high, very high), sensation of pain elsewhere, and analgesic consumption. Pain catastrophizing was assessed by means of the 13-item Pain Catastrophizing Scale (PCS), which measures thoughts and feelings when experiencing pain [18]. Elements of the PCS include scores of magnification, rumination, and helplessness. Together, these elements provide an overall score of catastrophizing (each item on a scale from 0 to 4, with a maximal score of 52).

2.2 Bedside sensory testing

The same two examiners (PLP, PB) carried out all the testing, which took place in a quiet room with a temperature of 22–23°C with the patient lying relaxed in a comfortable position. Brush-evoked allodynia was assessed by lightly stroking the thoracic wall using a brush (SENSELab™ Brush 0.5; Somedic AB, Hörby, Sweden), which was moved from outside the affected area along eight different lines converging towards the center. Pinprick hyperalgesia, pinprick hypoalgesia (a reference value for pinprick was obtained from a control area in the anterior part of the thorax), and cold allodynia were determined in the same manner using a a monofilament (Touch Test™, 60 g, Stoelting Co., Wood Dale, IL, USA) and a cold thermal roller at 20^oC (Somedic AB, Hörby, Sweden), respectively. If present, areas of spontaneous pain, brush- and cold-evoked allodynia, pinprick hyper- and hypoalgesia were mapped on the patients and transferred to a transparent paper, and the area was calculated in square centimetres. Intensities of evoked pain were recorded on the NRS. Scar lengths were also measured.

2.3 Surgical procedure, anesthesia, and pain management

Surgery was performed via bilateral thoracotomy. With the patient in a lateral decubitus position, a posterolateral thoracotomy was performed through the fifth intercostal space or alternatively intercostal space 6. Both the latissimus dorsi muscle and serratus anterior muscles were divided. The intercostal muscle was freed using electrocautery and a chest retractor was inserted. The recipient lung was removed and the donor lung was sewn in. First, the bronchial anastomosis was done on both the membranous wall and on the cartilaginous wall. Second, the pulmonary artery anastomosis was performed. Finally, the left atrial anastomosis was performed. Two 28 F chest tubes were inserted into the pleural cavity before closure. Afterwards, the patient was turned around and an identical procedure was performed on the other side of the thorax.

Prior to anesthesia, a thoracic epidural catheter was inserted at Th 5/6 or alternatively Th 6/7 with the loss of resistance technique. General anesthesia was then induced with fentanyl, midazolam, propofol, and cisatracurium and maintained with remifentanil, propofol, cisatracurium intravenously, and a continuous epidural infusion of bupivacaine with bolus doses as required. Postoperatively, pain was managed with an epidural infusion of bupivacaine for a maximum of 5 days and an oral regimen of modified-release paracetamol, extended-release oxycodone hydrochloride, and pregabalin, supplemented with oral or intravenous oxycodone as required.

2.4 Data analysis and statistics

Statistical analysis was performed using Stata version 12 (StataCorp, College Station, TX, USA) and SPSS version 20, Chicago, IL, USA). Data with normal distribution were presented as mean and standard deviation (SD). Data that did not follow normal distribution were presented as median and ranges. Scar lengths on sides with or without pain were compared with the Mann-Whitney test for unpaired data.

The χ² test for Hardy-Weinberg equilibrium was used to test for no intraindividual uniformity between the two sides as regards the presence or not of spontaneous pain and hyperphenomena (brush- and cold-evoked allodynia, pinprick hyperalgesia) and hypoalgesia. We calculated Spearman’s correlation coefficient for the within-patient comparison of side differences in reported pain intensity and area of evoked pain. Differences in the maximal area of evoked pain on the two sides of the thorax and differences in reported average pain intensity over the past week on the two sides of the thorax in the individual patient were used for calculation. A positive association would indicate an association between reported pain intensity and area of evoked pain.

The primary outcome was the presence of any pain (i.e. NRS>0) located to the thoracic scar area (bilateral, unilateral pain, or no pain). Secondary outcomes were the presence of moderate to severe pain located to the thoracic scar area (i.e. NRS≥3) and the presence of any hyperphenomena (i.e. either brush and/or cold allodynia and/or pinprick hyperalgesia). Also, the presence of each of the three hyperphenomena and the presence of hypoalgesia were calculated separately.

3.1 Patients

Twenty-two patients were approached for participation in the study. Two patients were not included: one did not understand Danish and one did not want to participate. Thus, 20 patients answered pain questionnaires and underwent bedside sensory testing 6–12 months after bilateral thoracotomy (see Table 1 for baseline characteristics). Lung transplantation was performed because of pulmonary fibrosis (n=6), alpha-1 antitrypsin deficiency (n=3), cystic fibrosis (n=3), chronic obstructive lung disease (n=3), lung emphysema (n=2), allergic alveolitis (n=1), childhood infectious lung disease syndrome (n=1), and sequelae following a gas explosion (n=1). Complications occurred in eight patients and included transplant rejection (n=3), diaphragmatic paralysis (n=2), infection (n=2, Epstein Barr virus/Aspergillosis), scar herniation (n=1), and reoperation for costal fracture with lung herniation (n=1). Four of the complications were unilateral. As can be seen from the descriptions below there seems to be no relation between complications and pain and hyperphenomena: one patient was re-operated on the right side because of costal fractures and herniation only on the right side, 10 months after he had severe pain bilaterally and hyperphenomena were similar on the two sides. One patient had a hernia on the left side (no reoperation), 6 months after she had no pain on either side of the thorax and similar objective findings on the two sides of the thorax. One patient had a right sided paresis of the diaphragm, 6 months postoperatively he had no pain on either side of the thorax, and minimal objective findings on both sides of the thorax. One patient had a left sided paresis of the diaphragm, 6 months after he had moderate pain on the left side but no hyperphenomena on this side.

3.2 Pain characteristics

Table 2 summarizes the pain characteristics. Nine patients reported pain on each side of the thorax, five patients had pain on one side only (all on the left side), and six patients had no pain (pain defined as average pain intensity over the past week, i.e. NRS>0). Statistical analysis (χ²-test for Hardy-Weinberg equilibrium) showed an intraindividual uniformity of pain between the two sides (p=0.029). The same uniformity in pain was found when analysing data from patients with spontaneous pain ≥3 on the NRS: 5 with bilateral pain, 4 with unilateral pain, and 11 with either no pain or pain <3 on the NRS (p=0.012).

Nine patients consumed analgesics daily: seven patients because of thoracic pain only and two patients because of thoracic pain and pain elsewhere. Eight patients took opioids and three took pregabalin. Thoracic pain affected daily activities much (n=4), some (n=1), little (n=4), and not at all (n=5) in the 14 patients who reported thoracic pain. PCS scores were generally low (Table 1).

Scar length was median 22 cm (range: 16–32 cm; n=2×20) and did not differ between sides with pain [median 21 cm (range: 16–32 cm; n=21)] compared to sides without pain [median 19 cm (range: 19–32 cm; n=19)], p=0.226.

3.3 Sensory testing

As can be seen from Table 3, most patients had positive sensory findings. On the right side of the thorax 18/20 patients had either sensory loss or gain (8/9 with pain). On the left side, 19/20 had either sensory loss or gain (13/14 with pain). Hyperphenomena were present on both sides of the thorax in 13 patients, on one side in four patients, and three patients had no hyperphenomena. Statistical analysis showed an intraindividual uniformity of hyperphenomena between the two sides (p=0.011). The same applied when analysing brush-evoked allodynia (p=0.037) and pinprick hyperalgesia (p=0.026) separately, but not cold-evoked allodynia (p=0.094) and pinprick hypoalgesia (p=0.094).

Within subjects, there was no correlation between the difference in average pain scores the past week between the sides of the thorax and the difference in the maximal area of hyperphenomena (brush- and cold-evoked allodynia and pinprick hyperalgesia) (Spearman’s rho=0.064, p=0.79) (Fig. 1) or the difference in the maximal evoked pain intensity (Spearman’s rho=−0.059, p=0.81) or the difference in the area of hypoalgesia (Spearman’s rho=−0.24, p=0.31), suggesting that within subjects there was no relation between patient-reported pain and evoked pain intensity or area of evoked pain or hypoalgesia.

Figure 1:

The relationship between differences in the maximal area of evoked pain on the two sides of the thorax (x-axis) and differences in reported average pain intensity over the past week on the two sides of the thorax (y-axis) for individual patients. Positive values indicate more pain/larger areas on the left side compared with the right side.